NekRS.
nekRS is an open-source, GPU-first spectral-element CFD solver from Argonne that solves the incompressible and low-Mach Navier–Stokes equations with scalar transport on high-order hexahedral elements, running through MPI and the OCCA portability layer on CUDA, HIP, SYCL and CPU backends. It adds LES and RANS models, moving meshes, overset grids, conjugate heat transfer and Lagrangian particles, targeting exascale machines.
- What are the resolved velocity, pressure and temperature fields in a turbulent coolant channel or blanket flow path, and what wall heat flux and pressure drop follow?
- How does conjugate heat transfer between a flowing fluid and the surrounding solid affect peak material temperatures?
- How does a thermal-hydraulic design scale to billions of degrees of freedom on GPU systems?
- Incompressible and low-Mach Navier–Stokes with scalar transport
- High-order curvilinear hexahedral spectral elements
- LES and RANS (k-tau, SST, DDES, IDDES)
- Arbitrary Lagrangian–Eulerian moving meshes; overset multi-session grids
- Conjugate fluid–solid heat transfer; Lagrangian particles
- GPU execution through OCCA (CUDA, HIP, SYCL) and CPU backends
Limitations
Linux and macOS only, built from source with C++17, MPI and CMake; no official binary or container distribution. Meshes must be conformal hexahedral. The project warns that user-facing APIs change between releases without backward compatibility, and releases are infrequent.
A general thermal-fluids CFD code, not fusion-specific. Its fusion relevance lies in blanket and coolant thermal-hydraulics: Argonne-led research papers report nekRS simulations of a fusion technology test facility and a spectral-element MHD formulation motivated by liquid-metal flows, but MHD is a research extension rather than a listed release feature. Fission use is the more established nuclear application.
No reference workflow documents this tool yet. Its evidence status is “Listed”: A reviewed catalog entry with upstream links.
Developed by Argonne National Laboratory (copyright UChicago Argonne, LLC), hosted under the Nek5000 GitHub organisation. No individual profiles are published for this project yet.
- Awesome Nuclear — Curated by Paul Romano and contributors
This entry was written by Morton Labs from public sources and last checked on 2026-10-10. Spotted an error, or want something changed or removed? Email zach@mortonlabs.ai with the page address.
Sources checked for this page (3)
- nekRS LICENSE — BSD-3-Clause-style licence with Argonne header
- nekRS v26.0 release — Latest release and date
- Crossref record for Fischer et al. 2022 — Citation details